++
This JAMA Guide to Statistics and Methods reviews the use of mediation analysis to evaluate possible mechanisms that the effects of interventions are presumed to work through.
++
In a study published in JAMA Network Open, Silverstein et al1 used mediation analysis to investigate how a problem-solving educational program prevented depressive symptoms in low-income mothers. Using data from a randomized trial, the authors tested 8 plausible mechanisms by which the intervention could have its effects. They concluded that problem-solving education reduced the risk of depressive symptoms in low-income mothers primarily by reducing maternal stress.
+++
Why Is Mediation Analysis Used?
++
The effects of health and medical interventions are often presumed to work through specific biological or psychosocial mechanisms. Possible mechanisms can be evaluated using mediation analysis.
+++
Description of Mediation Analysis
++
In mediation analysis, the effect of an intervention on an outcome is partitioned into indirect and direct effects. Indirect effects work through mediators of interest, whereas direct effects work through other mechanisms. These effects are often shown in a diagram (Figure 10). Mediation analysis can estimate indirect and direct effects and the proportion mediated, a statistical measure estimating how much of the total intervention effect works through a particular mediator.
++++
Two broad analytical approaches are used to conduct a mediation analysis: statistical and causal. Statistical mediation analysis uses regression models to estimate the strength of intervention-mediator and mediator-outcome effects. These regression coefficients can then be multiplied to estimate the indirect effect.2 Statistical mediation analysis is limited by its inability to accurately model situations in which there are nonlinear relationships between the intervention, mediator, and outcome or ...